In an implementation, a method for pixel level roadway crack segmentation is provided. The method includes: receiving a plurality of roadway range images; generating a plurality of image patches from the plurality of roadway range images; generating a crack map for the plurality of image patches by a DCNN; and generating a crack map for the plurality of roadway range images based on the generated crack map for the plurality of image patches.
Legal claims defining the scope of protection, as filed with the USPTO.
3. The method of claim 1, wherein generating the first plurality of 3D image patches from the first plurality of 3D roadway range images comprises generating the first plurality of 3D images patches without filtering or pre-processing.
7. The method of claim 6, wherein the encoder module comprises a plurality of 2D convolutional layers, and wherein each 2D convolutional layer decreases a height and a width of input features of a previous 2D convolutional layer by a factor of 2, and wherein output features of each 2D convolutional layer are processed through a Batch Normalization layer followed by a Leaky Rectified Linear Unit.
8. The method of claim 6, wherein the decoder module comprises a plurality of 2D transposed convolutional layers, and wherein each 2D transposed convolutional layer increases height and width dimensions of input features of a previous 2D transposed convolutional layer by a factor of 2, and wherein output features of each 2D transposed convolutional layer are processed through a Batch Normalization layer followed by a Leaky Rectified Linear Unit.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
May 6, 2021
December 3, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.